QUESTION: ... I have a series of 3-4 moderately sized studies with the data collected on different scanners (so it is not feasible to combine the raw data). I was wondering if you had any guidelines or feeling about the number of independent datasets that should go into an SDM analysis. Helpfully, I would have effect size/statistical maps for all datasets.

ANSWER: In "typical" SDM meta-analyses, i.e. mostly based on peak data and aiming to summarize all the studies in the field, I would suggest to include more studies. However, in your case I think that 3-4 studies may be enough, as a) you have all the data (not only peaks); and b) you are not aiming to summarize all the studies in the field, but only to combine your data.